Evaluation of Particle Swarm Optimization Algorithms for Weighted Max-Sat Problem: Technical Report

An experimental evaluation is conducted to asses the performance of 4 different Particle Swarm Optimization neighborhood structures in solving Max-Sat problem. The experiment has shown that none of the algorithms achieves statistically significant performance over the others under confidence level of 0.05.

Viewpoint Invariant Object Detector

Object Detection is the task of identifying the existence of an object class instance and locating it within an image. Difficulties in handling high intra-class variations constitute major obstacles to achieving high performance on standard benchmark datasets (scale, viewpoint, lighting conditions and orientation variations provide good examples). Suggested model aims at providing more robustness to detecting objects suffering severe distortion due to < 60{\deg} viewpoint changes. In addition, several model computational bottlenecks have been resolved leading to a significant increase in the model performance (speed and space) without compromising the resulting accuracy. Finally, we produced two illustrative applications showing the potential of the object detection technology being deployed in real life applications; namely content-based image search and content-based video search.